[Bug 752613] qtdemux: unsigned raw 16 bit PCM audio not working

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=752613

--- Comment #18 from manasa.athreya at lge.com ---
(In reply to manasa.athreya from comment #16)
> (In reply to Tim-Philipp Müller from comment #15)
> > > The following is the case observed:
> > > 
> > > 1. In NONE/raw case the code is not checking the type of endian, which is
> > > simply set to Little endian format.
> > 
> > I don't understand what you mean by this. With the current code, no
> > endianness is set for NONE/raw, and it's not needed since it's only U8. With
> > your patch it's set to BIG_ENDIAN, which is correct according to the
> > quicktime specs. With my suggestion to fallthrough to the 'twos' case, it
> > would also be set to big endian.
> > 
>   Since initially break statement is place after NONE/raw, case was not
> going   to twos case so we placed code again in NONE/raw.
> >  
> > > 2. The attached file has 16-bit PCM data which is being set to 8-bit data by
> > > default.
> > 
> > I understand this. And your patch fixes it. I just asked you to fix it
> > differently, with less code.
> > 
> Yes, Will try to fix it in less code.
> >  
> > > 3. stream->bytes_per_packet might not be always 1 in NONE/raw case.
> > 
> > I understand this. And your patch fixes it. I just asked you to fix it
> > differently, with less code.
> > 
> > Also:
> > 
> > 4. In the 'twos' case, the data might be 8-bit (S8 in this case) as well.
> > The code currently doesn't handle that.
> > 
> > I think the NONE/raw/twos case could be unified into one block.
> 
> I agree to this idea of unifying.
> 
> Also:
> I will try to unify and submit the same.

I have unified NONE/raw/twos case .Please check the attached patch.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.